摘要
针对传统粒子滤波面临的重要密度函数的选取和粒子多样性丧失引起的样本贫化问题,提出基于无迹卡尔曼滤波和权值优化的改进粒子滤波算法。与传统的粒子滤波算法相比,有两点改进:首先该算法采取无迹卡尔曼滤波产生建议分布函数;其次,在重采样过程,提出基于权值优化的改进重采样算法来增加粒子的多样性。仿真结果表明,改进算法降低了粒子滤波算法的粒子退化程度并避免样本贫化现象的出现,更加接近真实值,提高了跟踪精度。
Aiming at the problems of the importance function choice and the sample impoverishment after resampling,an improved particle filter algorithm was proposed in this paper,which based on the unscented Kalman filter and weight optimization.Compared with the traditional particle filter,this algorithm had two improvements,UKF was used to generate the importance density function,and weight optimization was used to ensure all useful information inherited,which could maintain the diversity of particle.The theory analysis and simulation showed that the improved particle filter algorithm could solve particle degeneracy and avoid sample impoverishment.
引文
[1]吴兆平,朱凯然,苏涛,等.采用改进粒子滤波的雷达扩展目标检测前跟踪[J].西安电子科技大学学报,2011,38(2):99-104.
[2]李浩,王润亮,彭华.粒子滤波盲均衡译码联合算法[J].西安电子科技大学学报,2013,40(1):123-128.
[3]胡洪涛,敬忠良,李安平,等.非高斯条件下基于粒子滤波的目标跟踪[J].上海交通大学学报,2004,38(12):1996-1999.
[4]周明,涂宏斌.一种基于改进粒子滤波的多目标检测与跟踪方法[J].华东交通大学学报,2016,33(2):121-126.
[5]孙景乐,唐林波,赵保军,等.基于瑞丽分布的粒子滤波跟踪算法[J].电子信息学报,2013,35(4):763-769.
[6]Zou J Y,Wu C H,Ma T G.Heterogeneous Integrated Beam-switching/Retrodirective Array Using Synthesized Transmission Lines[J].IEEE Transactions on Microwave Theory and Techniques,2013,61(8):3128-3139.
[7]Gordon N J,Salmond D J,Smith A.A novel approach to nonlinear and non-Gaussian Bayesian state estimation[C]//IEEE Proceeding,1993,140:107-113.
[8]LIU J S,CHEN R.Sequential monte carlo methods for dynamic systems[J].J.of the American Statistical Association,1995,90(2):567-576.
[9]DOUCET A,GODSILL S.On sequential monte carlo sampling methods for bayesian filtering[R].Cambridge:University of Cambridge,1998.
[10]TAI S.Cloud service engineering[C]//Proc of the 18th IEEE International Workshops on Enabling Technologies:Infrastructures for Collaborative Enterprises.Washington DC:IEEE Computer Society,2009:3-4.
[11]Wan E A,Van Der Merwe R.The unscented kalam filter for nonlinear estimation[C]//Proceeding of the IEEE2000Adaptive Systems for Signal Processing,Communications,and Control Symposium.Piscataway:IEEE,2000:153-158.
[12]ARULAMPPALAM M S,MASKELLS,GORDON N,et al.A tutorial on particle filter for online nonlinear/non gaussssian bayesian tracking[J].IEEE Trans.on Signal Processing,2002,50(2):174-188.
[13]杜正聪,辛强,邓寻,等.基于权值优化的粒子滤波算法研究[J].重庆师范大学学报,2015,32(3):124-129.
[14]朱志宇.粒子滤波算法及其应用[M].北京:科学出版社,2010.
[15]朱军.基于权值优化的粒子滤波算法研究及其在目标跟踪中的应用[D].安徽:安徽建筑大学,2014:1-61.